Benzothiophene Selective Estrogen Receptor Modulators with

May 10, 2007 - Tripos force field and Gasteiger-Huckel charges,47,48 relative energies were calculated ...... Mathematical Programming 1977, 12,. 241-...
0 downloads 0 Views 341KB Size
Download PDF
2682

J. Med. Chem. 2007, 50, 2682-2692

Benzothiophene Selective Estrogen Receptor Modulators with Modulated Oxidative Activity and Receptor Affinity Zhihui Qin, Irida Kastrati, R. Esala P. Chandrasena, Hong Liu, Ping Yao, Pavel A. Petukhov, Judy L. Bolton, and Gregory R. J. Thatcher* Department of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, UniVersity of Illinois at Chicago, 833 South Wood Street, Chicago, Illinois 60612-7231 ReceiVed January 19, 2007

The regulation of estrogenic and antiestrogenic effects of selective estrogen receptor modulators (SERMs) is thought to underlie their clinical use. Most SERMs are polyaromatic phenols susceptible to oxidative metabolism to quinoids, which are proposed to be genotoxic. Conversely, the redox reactivity of SERMs may contribute to antioxidant and chemopreventive mechanisms, providing a new approach to improve the therapeutic properties of SERMs. An improved synthetic strategy was developed to generate a family of benzothiophene SERMs. Using computational modeling methods and measurements of antioxidant activity and estrogen receptor (ER) ligand binding, this SERM family was shown to provide both a range of ERR/ ERβ selectivity from 1.2- to 67-fold and a range of redox activity. Antioxidant activity was successfully modulated by varying a substituent remote from the OH group; the source of the antioxidant capacity. An efficient synthetic procedure is reported yielding benzothiophene SERMs wherein redox activity and ER affinity are modulated. Introduction Tamoxifen is the archetypal selective estrogen receptor modulator (SERMa). Despite the demonstrated, increased risk of endometrial cancer, tamoxifen has been the therapy of choice in the endocrine treatment of all stages of hormone-dependent breast cancer and in the primary and secondary chemoprevention of breast cancer.1 Although the introduction of aromatase inhibitors may change this clinical paradigm, SERMs are likely to be in clinical use for many years.2 The increased use of SERMs is anticipated on the basis of favorable clinical trial results for the benzothiophene SERMs, raloxifene and arzoxifene, and because SERMs are hoped to provide an alternative to current hormone replacement therapy (HRT) that has been causally linked to breast cancer.3 Raloxifene is in current clinical use in post-menopausal osteoporosis and is expected to find use in other postmenopausal indications associated with HRT.4,5 The STAR trial (study of tamoxifen and raloxifene) reported that raloxifene was as effective as tamoxifen in breast cancer chemoprevention in postmenopausal women at high risk and was less likely to cause the potentially dangerous side effects associated with tamoxifen, such as uterine cancer and blood clots. The RUTH trial (raloxifene use for the heart) did not show a significantly increased risk of coronary artery disease, although there is still debate on the potential beneficial or negative effects of raloxifene on other cardiovascular events.5-7 Arzoxifene, designed to improve upon the therapeutic properties of raloxifene, is in late stage clinical trials with the promise of substantial therapeutic benefits and is likely to find use in cancer chemoprevention.8,9 The carcinogenic effects of tamoxifen have been attributed variously to regulation of gene transcription (i.e., hormonal carcinogenesis) and to genotoxicity due to oxidative metabolites * To whom correspondence should be addressed. Tel.: 312-355-5282. Fax: 312 996 7107. E-mail: [email protected]. a Abbreviations: BDE, bond dissociation energy; DMA, desmethylarzoxifene; X-DMA, 4′-X-4′-desmethoxyarzoxifene; DPPH, diphenylpicrylhydrazyl radical; ER, estrogen receptor; HRT, hormone replacement therapy; LBD, ligand binding domain; SERM, selective estrogen receptor modulator.

(i.e., chemical carcinogenesis). Chemical carcinogenesis can contribute to cancer initiation through damage to DNA and other biomolecules following drug bioactivation to redox-active and electrophilic quinoid metabolites (o-quinones, quinone methides, and di-quinone methides).10,11 Human estrogens and equine estrogens contained in current HRT agents are also proposed to elicit hormonal and chemical carcinogenesis pathways, the latter via o-quinone metabolites.12,13 Interestingly, many SERMs in clinical use and clinical development are also highly susceptible to oxidative metabolism to electrophilic and redoxactive quinoids simply because they are based on polyaromatic phenol scaffolds.14 The SERMs, raloxifene, desmethylarzoxifene (DMA), acolbifene, toremifene, and droloxifene are all oxidatively metabolized to quinoids, which have been shown to form adducts with biomolecules, including glutathione (GSH), proteins, and nucleosides (Scheme 1).14-20 Whereas generation of reactive oxygen species (ROS) and covalent modification of biomolecules by redox-active quinoids may contribute to initiation and promotion of carcinogenesis, induction of oxidative stress and oxidation or covalent modification of sensor proteins may trigger cellular stress responses that are cytoprotective.21,22 This balance between the carcinogenic and the chemopreventive capacity of a drug is determined by the reactivity toward oxidative bioactivation and the chemistry of the reactive metabolite formed and, therefore, can be controlled by structural modification. The contribution of oxidative bioactivation to therapeutic activity versus toxicity is of particular relevance to SERMs, which are designed for chronic use in healthy women who are peri- or postmenopausal or who have known risk factors. The continued development of SERMs based on polyaromatic phenolic scaffolds requires increased understanding of the influence of oxidative bioactivation. Intensive research is currently directed at discovery of the “ideal SERM”: an agent that is antiestrogenic in breast and endometrial tissue, but proestrogenic in the vasculature and brain, which would be of use in cancer chemoprevention and an attractive alternative to HRT. However, there has been little

10.1021/jm070079j CCC: $37.00 © 2007 American Chemical Society Published on Web 05/10/2007

Benzothiophene SERMs

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2683

Scheme 1

Scheme 2 a

a Reagents and conditions: (a) KOH, EtOH, EtOAc; (b) PPA, 130 °C; (c) CH C(O)NHBr; (d) H O , TFA, DCM; (e) NaH, C H N(CH ) OC H OH, 3 2 2 5 10 2 2 6 4 DMF; (f) (i) HCl, ether; (ii) TMSCl, Ph3P, THF, reflux; (g) (i) HCl, ether; (ii) BF3SMe2; (h) LiAlH4, THF, reflux; (i) (i) HCl, ether; (ii) BF3SMe2.

Scheme 3 a

a Reagents and conditions: (a) NaOMe, cat. AcOEt, CuI, DMF, MeOH, reflux; (b) L-proline, NaOH, CH3SO2Na, CuI, DMSO.

attention to structural modifications designed to control oxidative bioactivation and thereby enhance therapeutic activity and attenuate toxicity.23 This approach requires a family of SERMs in which structure is used to modulate both redox reactivity and activity at the estrogen receptor. To that end, a family of benzothiophene SERMs related to arzoxifene has been synthesized, requiring development of a new synthetic methodology for arzoxifene itself.

Results and Discussion Arzoxifene (1; Scheme 1) is a structural analogue of raloxifene in which the carbonyl hinge has been replaced by an ether linkage and the 4′-hydroxy group is methylated. Arzoxifene is in late stage clinical trials as a next generation SERM with promise of substantial therapeutic benefits24 that are suggested to result from (a) increased antiestrogenic potency and (b) improved bioavailability relative to raloxifene.25 DMA (2; Scheme 1) is an active metabolite of arzoxifene, which has been observed with highly variable steady-state plasma concentrations.8 In vitro metabolic studies showed that both DMA and raloxifene undergo bioactivation to electrophilic diquinone methides (Figure 1), resulting in potentially cytotoxic actions: depletion of cellular GSH, irreversible inhibition of P450 3A4, and liver protein modification.15,16,19,20 The desired chemopreventive actions of SERMs will be compromised by the formation of covalent adducts between electrophilic quinoid metabolites and cellular proteins or DNA if these adducts cause genotoxicity or organ toxicity. In an effort to obtain safer benzothiophene SERMs that retain efficacy and have attenuated reactivity toward bioactivation, the arzoxifene analogue, 4′-fluoro-4′-desmethoxyarzoxifene, has been developed (F-DMA, 3, Scheme 1). F-DMA showed similar antiestrogenic activity to both DMA and raloxifene, and 4′-fluorination was shown to successfully block the formation of an electrophilic diquinone methide (Figure 1) and to suppress phase II metabolism. These properties are predictive of improved bioavailability compared to DMA and raloxifene.15,16 Drug Design Rationale. Appropriate structural modifications designed to minimize drug bioactivation are sometimes incor-

2684

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11

Qin et al.

Figure 3. (A) Docking poses for the synthetic analogs in ERR/LBD. (B) Superposition of the rigid (green) vs relaxed (cyan) binding cavity of ERR/LBD with docked 4′-XDMA (X ) NH(CH2)2Cl). Figure 1. Bioactivation of raloxifene and DMA is blocked for F-DMA.

Figure 4. analogues.

Figure 2. Plots of EHOMO (squares) and relative BDE (triangles) vs Hammett parameter for benzothiophene derivatives from AM1//B3LYP/ 6-31+G* calculations: open circles represent 4-substituted compounds selected for synthesis.

porated in the lead optimization stage of drug discovery,23 but most SERMs retain the polyaromatic phenolic core that is susceptible to oxidative bioactiviation. Although minimizing drug bioactivation is often thought to reduce the risks associated with toxic metabolites, bioactivation to a benign redox-active metabolite that induces oxidative stress and oxidation or covalent modification of sensor proteins may trigger cytoprotective cellular responses that contribute to chemoprevention. Keap1 is a notable example of a sensor protein that responds to redoxactive compounds to mediate induction of phase II enzymes.21,22 An additional potential benefit is antioxidant activity, which has been proposed to contribute to SERM biological activity.26,27 The potential negative outcomes of oxidative bioactivation include modification of liver proteins, drug-drug interactions, and liver dysfunction. Further complicating this picture is the proposal that quinoid metabolites of estrogenic compounds are ligand-independent ER modulators.28 To explore the influence of bioactivation on SERM activity and toxicity and hence improve SERM design, a homologous family of SERMs was required with modulated ER binding and redox reactivity. A benzothiophene core was selected because of the importance of raloxifene and arzoxifene and on the basis of the promising preliminary studies with F-DMA. Naturally, some, but not all, of the benzothiophene SERMs selected for synthesis have been described in patents but, with the exception of arzoxifene itself, there is little or no data published on biological activity and reactivity.

Retro synthetic analysis of 4′-substituted arzoxifene

I. Redox Activity. Antioxidant capacity can be quantified by the ArO-H bond dissociation energy (BDE), and the relative oxidative capacity for phase II enzyme induction has been correlated with EHOMO.29 Therefore, calculations were performed to predict antioxidant capacity at the AM1//B3LYP/6-31+G* level on a series of 4′-substituted benzothiophenes (Figure 2; X ) NH2, OMe, OH, NHC(O)CH3, Me, H, F, Br, CHO, CN, SO2Me, NO2). Linear correlations were observed between standard Hammett parameters and both EHOMO and BDE (r2 ) 0.98 and 0.92, respectively; Figure 2). On this basis, to provide a range of redox reactivity, the 4′-substituted-4′-desmethoxyarzoxifenes (X-DMA series) selected for synthesis were X ) NH2, OMe, OH, H, F, Br, SO2Me, and NHR (R ) alkyl, acyl). This family contains, arzoxifene, DMA, and F-DMA, in addition to interesting compounds such as a methylsulfonyl derivative that in simile with the non-CNS penetrating SERM, LY2066948,30 are predicted to have reduced brain bioavailability and hence diminished ovarian stimulation via actions in the hypothalamus.31 II. ER Binding. To assist in drug design, the coordinates for the ERR and ERβ ligand binding domains (LBD) were extracted from the crystal data of the raloxifene-ERR/LBD complex (PDB code: 1ERR) and tamoxifen-ERβ complex (PDB code: 2FSZ), respectively. The ERR/LBD structure indicated that the 4′-OH is involved in a hydrogen-bonding network that includes His-524, with which it forms a hydrogen bond.32 The simplistic prediction would be that the loss of this hydrogen bond in derivatives such as H-DMA (12) would reduce ER affinity. Because His-524 could potentially act as a nucleophile toward an electrophilic 4′-substituent with potential to covalently modify the ERR ligand binding site, for X ) NHR, R ) (CH2)2Cl (18) and C(O)CH2Cl (16) were selected. A chloromethyl ketone group is known to alkylate histidine residues,33 and a nitrogen mustard chloroethylamine group is known to alkylate via an aziridine intermediate.34

Benzothiophene SERMs

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2685

Scheme 4 a

a Reagents and conditions: (a) NaN , CuI, L-proline, NaOH, DMSO, EtOH; (b) (i) HCl, ether; (ii) BF SMe ; (c) ClCH CHO, NaBH CN, HCl, MeOH; 3 3 2 2 3 (d) ClCH2COCl, pyridine, CH2Cl2; (e) (i) HCl, ether; (ii) BF3SMe2, 70 °C.

Raloxifene binding in ERR/LBD is achieved through a combination of specific hydrogen bonds and complementarity of the binding cavity with the nonpolar portions of the ligand. Given the similarity between raloxifene and the benzothiophene SERMs described in this study, docking to ERβ/LBD was anticipated to provide good predictive binding data. Whether FlexX was run on its own or with FlexPharm constraints and/ or more stringent docking criteria, the bulkier 4′-substituents (X ) SO2CH3, NH(CH2)2Cl, NHC(O)CH2Cl, and in some cases OCH3) failed to position inside the LBD cavity. In FlexX, X-DMA derivatives with smaller 4′-substituents positioned inside the LBD, in simile with the raloxifene-ERR/LBD crystal structure. (Figure 3A). The prediction from the preliminary computational studies was that the selected family of X-DMA derivatives would provide a spectrum of ER affinity and redox reactivity. Synthesis. It was necessary to develop a new synthetic methodology toward arzoxifene and its analogues. The published synthesis of arzoxifene is multistep and is not readily amenable to adaptation to generate arzoxifene analogs.35 In addition, the starting material for the published 11-step synthesis of arzoxifene, 1-methanesulfonyloxy-4-bromobenzene, is not commercially available. A new synthesis was designed to provide a versatile common intermediate for preparation of arzoxifene and novel arzoxifene analogs. The 4′-bromo derivative, 4, provides an ideal synthetic intermediate because it is amenable to copper(I) iodide-catalyzed aryl bromide derivatization (Figure 4). The synthesis of 4 was proceeded by the adaptation of methods developed by us for F-DMA (Scheme 2).15 3-Methoxybenzenethiol and 2,4′-dibromoacetophenone coupling under basic conditions gave the β-ketosulfide 5, followed by cyclization and rearrangement at 130 °C in polyphosphoric acid (PPA) to yield a mixture of rearranged and isomeric products. The desired rearranged product 6 was readily separated in good yield by simple ethyl acetate extraction and filtration. Recovered unrearranged reactant was subjected to subsequent rearragement to readily provide increased quantities of 6. Bromination of 6 with bromoacetamide in quantitative yield followed by oxidation of the 3-bromobenzo[b]thiophene 7 with H2O2 gave the corre-

sponding sulfoxide 8, in which the bromide was activated by the electron-withdrawing sulfoxide toward SNAr reaction.36 The subsequent phenol displacement proceeded smoothly, and compound 9 was obtained in high yield. The desired 4′-bromo synthetic intermediate, 10, was obtained by reduction with Ph3P/ TMSCl,37 requiring prior protonation of the piperidine nitrogen by acidification. LiAlH4 reduction at sulfur, employed in the preparation of DMA25 and 4′-F-DMA,15 led to substitution of the 4′-bromide with hydride and provided a useful synthetic route to desmethoxyarzoxifene, 12, from 9. Methyl ether deprotection of 10 and 11 using BF3·SMe238 gave the 4′-bromo4′-desmethoxy and 4′-desmethoxy arzoxifene analogues 4 and 12, respectively (attenuation of nitrogen basicity by acidification was also required prior to use of the BF3 complex). Arzoxifene itself (1) was obtained by the single-step methanolysis of 4, effected by copper(I) iodide-catalyzed aryl bromide substitution with concentrated sodium methoxide solution and a catalytic amount of ethyl acetate that functions to prevent precipitation of the copper species during the reaction (Scheme 3).39 This modified arzoxifene synthesis reduces the overall number of steps compared to the published 11-step synthesis, uses low-priced commercial starting materials, and is amenable to scale-up. A sulfone-substituted arzoxifene analogue was readily prepared from 4 by direct introduction of the sulfone at the 4′position by coupling with sodium methanesulfinate under the catalysis of a copper(I) iodide/L-proline sodium salt, a methodology that has been recently reported (Scheme 3).40 The Cu(I) iodide/L-proline catalytic system can also be used in the synthesis of aryl azides by the coupling reaction of the aryl halide with sodium azide.41 This is an attractive route to 4′amino-4′-desmethoxyarzoxifene, an analogue that is predicted to possess similar estrogenic binding and oxidative bioactivation properties to DMA. Under catalysis of Cu(I) iodide/L-proline, the reaction of compound 10 with sodium azide was achieved by switching the solvent system from the reported EtOH/H2O to a mixture of DMSO/ethanol heated at around 110 °C (Scheme 4). Instead of the azido product, the reaction directly gave the desired 4′-amino compound 14 as a single product in good yield, which might be due to the instability of the aryl azide at high

2686

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11

Figure 5. (A) Time plot of relative absorbance (515 nm) showing decay of DPPH radical in the presence of selected X-DMA SERMs. (B) Observed rate constants for quenching of DPPH radical in methanolic solutions by X-DMA derivatives plotted against EHOMO.

temperature (Scheme 4). The subsequent demethylation gave the product 15. Selective acylation on the introduced amino group gave the R-chloroacetylated product 16. The aniline mustard 18 was prepared by reductive amination reaction of 14,42 and BCl3·SMe2 complex43 was used to avoid the possible halogen exchange that might occur between BF3 and the chloride of the mustard. Antioxidant Activity. The simple 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay was run to measure the relative antioxidant capacity of the X-DMA series compounds and raloxifene. DPPH is a stable nitrogen radical, the scavenging of which can be measured by its decay at 515 nm (Figure 5). Although scavenging of stable nitrogen radicals is not a major function of antioxidants in vivo, the assay is routinely used to quantify antioxidant capacity and is useful for comparisons within structural families. As shown in Figure 5 for the monophenolic SERMs, an excellent correlation was seen between the rate of radical scavenging and the calculated EHOMO for the X-DMA series, validating the computational

Qin et al.

predictions. The observed pseudo first-order rate constants are shown in Table 1. DMA (2) and NH2-DMA (5) are not expected to follow a similar pattern to the monophenolic SERMs because these are able to form quinoids, a diquinone methide and quinone imine, respectively, and theoretically to quench 2 equiv of DPPH. Estrogen Receptor Binding Assays. Arzoxifene and its analogues were assayed in the standard ER competitive radioligand binding assay, using full length human recombinant ERR and ERβ and compared to raloxifene.44 As shown in Table 1, X-DMA ligands bind to ERR with affinities ranging from the most potent ligand 2 (DMA) to 16, which is a very poor ligand for ERR. All ligands are nonselective or selective for ERR over ERβ, owing to the larger ligand binding cavity of ERR, hence X-DMA ligands 13 and 18 are highly selective for ERR. Docking Studies. X-DMA ligands with smaller 4′-substituents docked into the rigid LBD binding sites, using FlexX, giving good correlations between D-Score and the experimental pIC50 (exp ∆G; r2 ) 0.912 for ERR and r2 ) 0.744 for ERβ). Of the scoring methods tested (G_Score, PMF_Score, D_Score, ChemScore, Total-Score), D_Score consistently showed the best correlation for FlexX-derived binding poses with both ERR and ERβ. Ligands with bulkier 4′-substituents (X ) SO2CH3, NH(CH2)2Cl, and NHC(O)CH2Cl and, in some poses, OCH3) failed to dock inside the rigid LBD cavity, but three of these were observed experimentally to be reasonable ligands for ERR. To account for this observation, docking was achieved using the DMA binding pose (Figure 3) as a starting point, followed by force field minimization of the protein-ligand complex. Using this method to examine the X-DMA ligands, D-Score again provided the best correlation. This scoring function, drawn from the molecular docking program DOCK, is a classical force field energy function, which sums van der Waals and electrostatic interactions in the ligand binding complex.45 This result is compatible with the dominant contributions to ER/LBD binding from van der Waals interactions with residues, such as Leu384 and Met421, and electrostatic interactions with (i) Arg394/Glu353,42 His524, and (iii) Asp351(303), leading to displacement of helix12 (Figure 6). Thus, computational docking was able to account for experimental ligand binding using a rigid receptor for ligands with smaller 4′-substituents and using an iteratively relaxed receptor cavity for the more sterically encumbered ligands. Using the Powell method46 for energy minimization with the Tripos force field and Gasteiger-Huckel charges,47,48 relative energies were calculated for binding of raloxifene and the X-DMA ligands to ERR and compared to the experimental binding energies, giving an excellent correlation for the ligands binding to the rigid receptor (Figure 7) but not for those that required relaxation of the receptor residues. This observation demonstrates the limitations of the crystal structure docking protocol for prediction of ligand binding and emphasizes that conformational distortion of the receptor in the region of the 4′-DMA position to accommodate larger ligands does not result in substantial loss of binding affinity. The crystal structure reveals that the 4′-substituent interacts with a less-ordered portion of the LBD. This observation suggests that further modification at the 4′-position can be explored to optimize ERR binding while ablating ERβ affinity. Crystal structures show two distinct conformations of the His-524 imidazole ring.32 Despite the indicated proximity of His-524 in the relaxed receptor docking and the good ERR affinity of 18, the present data do not provide any evidence for covalent modification of ERR. Conversely, the good selectivity of 18 for ERR/ERβ

Benzothiophene SERMs

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2687

Table 1. ER Binding Data and DPPH Radical Scavenging Rates for X-DMA SERMs Compared to Raloxifenea ER RBAb

ER IC50 (nM)

cmpds (X)

ER-R

ER-β

selectivity R/β

ER-R

ER-β

raloxifene 1 (OCH3) arzoxifene 2 (OH) DMA 3 (F) F-DMA 4 (Br) 12 (H) 13 (SO2CH3) 15 (NH2) 16 (NHC(O)CH2Cl) 18 (NH(CH2)2Cl)

0.91 ( 0.12 0.94 ( 0.29 2.51 ( 0.61 1.07 ( 0.04 0.70 ( 0.12 1.71 ( 0.23 0.71 ( 0.12 0.96 ( 0.08 0.03 ( 0.01 0.56 ( 0.15

0.03 ( 0.01 0.25 ( 0.01 1.77 ( 0.35 0.68 ( 0.26 0.25 ( 0.03 1.00 ( 0.05 0.009 ( 0.0005 0.34 ( 0.01 0.007 50

20.6 ( 2.7 21.5 ( 6.5 7.8 ( 1.9 17.2 ( 0.6 27.0 ( 4.6 10.9 ( 1.5 27.0 ( 4.8 19.3 ( 1.7 666 ( 219 35.6 ( 10

557 ( 146 66.3 ( 3.1 9.6 ( 1.9 27.9 ( 11 66.7 ( 8.5 16.3 ( 0.8 1800 ( 100 48.6 ( 1.9 2360 ( 18 >1600

DPPH scavenging 103 × kobs, s-1 4.0 ( 0.6 35 ( 7 9.3 ( 0. 4 6.0 ( 0.9 6.5 ( 0.3 8.9 ( 0.9 3.0 ( 0.5 220 ( 0.4 c c

Data shown is the mean ( S.D. for at least triplicate measurements. RBA values calculated relative to IC50 assayed for E2 control (RBA ) 1.0). c Not measured. a

b

Figure 6. Raloxifene (red) and DMA docked in the ERR/LBD showing key residues. The MOLCAD multichannel surface was generated in SYBYL. Distances between X-DMA ligands and key ERR/LBD residues, Å, are shown in the table below.

Figure 7. Correlation between experimental pIC50 (exp ∆G) and calculated ∆G binding (r2 ) 0.93).

further supports modification of the 4′ position of benzothiophene SERMs to enhance selectivity. Recently, a second low-affinity binding site for tamoxifen has been reported in a crystal structure of 4-hydroxytamoxifen-

ERβ/LBD (PBD ID: 2FSZ), therefore, it was of interest to screen the X-DMA ligands for binding.49 Most interactions at the low affinity binding site are van der Waals contacts provided by the amino acids within a radius of 5.5 Å from tamoxifen, with core subpocket residues including Leu306, Met309, Ile310 Val 328, Leu331, Glu332, and Trp335 (using ERβ 2FSZ reference numbering). The weak, nonspecific ligand association in this hydrophobic grove permits numerous binding poses for the X-DMA ligands, which would require probing by long-range molecular dynamics. The low affinity binding site is not proposed to inhibit estrogen binding to ER, but rather to antagonize coregulator binding, thus, binding of bulkier X-DMA ligands to this site cannot account for the competitive binding assay data shown in Table 1. Comparisons of X-DMA Activity/Reactivity. Except 16 and 18, the family of benzothiophene SERMs reported can be subdivided into those that are readily able to form quinoids (2, 15) and the subset that are anticipated to form semiquinones on oxidation (1, 3, 4, 12, 13). For the latter SERMs, two-electron oxidation is blocked by the 4′-modification, thus it might be

2688

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11

anticipated that these will undergo oxidation to o-quinones, an alternate pathway that is not blocked. In the latter subset, the redox activity is controlled by the remote 4′-substituent, which is expected to influence the rate of formation of o-quinones. For example, 13 is 10-fold less redox active than 1 (Table 1). The range of EHOMO values calculated for the SERM family is similar to that for families of compounds that have shown a wide range in chemopreventive index.29 However, the anticipated metabolism of both 1 and 12 to yield DMA (2) would complicate interpretation in more complex systems. Nevertheless, the data presented herein presents some interesting comparisons. For example, 4 and 13 have similar affinity for ERR, and both have attenuated redox activity because of electron-withdrawing 4′-substituents, but 13 is highly selective for ERR/ERβ, and 2 and 12 are comparably antiestrogenic, but only 2 can form a diquinone methide. The extrapolation of the spectrum of antiestrogenic and redox activity, measured in simple models herein, to complex cellular and in vivo systems awaits the completion of such studies measuring oxidative metabolism, Phase II enzyme induction, and estrogenic versus antiestrogenic endpoints. Conclusions An ideal SERM, antiestrogenic in breast and endometrial tissue, but proestrogenic in the vasculature and brain would be of use in cancer chemoprevention and as an alternative to hormone replacement therapy. The approaches taken have focused largely on amplifying antiestrogenic activity; screening families of SERMs for compounds with increased tissue selectivity and, more recently, isoform selectivity. There has been relatively little research directed at exploring the contribution of the redox reactivity and bioactivation that is common to many SERMs and to test the hypothesis that structural modification of SERMs to modulate bioactivation will lead to an improved SERM. The synthesis presented herein provides access to a structurally conservative family of SERMs related to the clinically important benzothiophene SERMs that are shown to manifest a spectrum of redox reactivity and ER ligand properties. Experimental Section Synthesis. 1H and 13C NMR spectra were obtained with Bruker Ultrashield 400 or Advance 300 spectrometer. Chemical shifts are reported as δ values in parts per million (ppm) relative to tetramethylsilane (TMS) for all recorded NMR spectra. Lowresolution mass spectra were recorded on a Agilent 1100 series LC/MSD ion trap instrument, using APCI as ionization method. High-resolution mass spectra were taken on a Micromass QTOF mass spectrometer, using ESI as the ionization method. All reagents and solvents were obtained commercially from Acros, Aldrich, and Fluka and were used without purification. 1-(4-Bromophenyl)-2-(3-methoxylphenylsulfanyl)ethanone (5). KOH (5.2 g, 87%, 80 mmol) was dissolved in ethanol (200 mL), and 3-methoxy benzenethiol (9.7 mL, 79 mmol) was added and stirred for 10 min. Ethyl acetate (80 mL) was added to this solution, then 2,4′-dibromoacetophenone (20 g, 72 mmol) was added in portions. Another 60 mL of ethyl acetate was added to dissolve some precipitate that was produced during the reaction after 3 h. The reaction mixture was stirred overnight at room temperature. Most solvent was removed under reduced pressure, the residue was partitioned between ethyl acetate and brine, and the organic phase was separated and dried over anhydrous MgSO4. Concentration and recrystallization using ethyl acetate gave desired compound 5 as a yellow solid (21.5 g, 89%). 1H NMR (acetone-d6, 400 MHz): δ 7.96-7.98 (m, 2H), 7.70-7.73 (m, 2H), 7.19-7.23 (m, 1H), 6.936.95 (m, 2H), 6.76-6.79 (m, 1H), 4.53 (s, 2H), 3.77 (s, 3H). 13C

Qin et al.

NMR (acetone-d6, 100 MHz): δ 193.9, 160.9, 137.5, 135.6, 132.7, 131.3, 130.7, 128.6, 122.1, 115.3, 113.0, 55.5, 40.8. APCI-MS: m/z 339.0/337.0 (100/98%) [M + H]+. 6-Methoxy-2-(4-bromophenyl)benzo[b]thiophene (6). PPA (70 g) was added to a 250 mL flask and heated to 80 °C with stirring (keep the speed of stirring as fast as possible). Compound 5 (10 g, 29.8 mmol) was added portionwise within 30 min, then the temperature of oil bath was elevated to 130 °C, and the reaction mixture was heated with stirring for 6 h. The reaction mixture was poured into 500 mL of rapidly stirring ice water to allow the PPA to be hydrolyzed. After 2 h, 150 mL of ethyl acetate was added and stirred for 20 min, and the crude product was collected by filtration and washed with 15 mL of H2O three times. The obtained brown solid was air-dried overnight to get the title compound, which was pure enough for the next step (2.9 g, 30%). The ethyl acetate solution containing unrearranged reactant was washed with H2O and NaHCO3 solution, concentrated, passed through a short silica gel column, and subjected to another rearrangement reaction to gave more desired product. An analytic sample of 6 was obtained by recrystallization using ethyl acetate. 1H NMR (DMSO-d6, 400 MHz): δ 7.81 (s, 1H), 7.73 (d, 1H, J ) 8.7 Hz), 7.63-7.68 (m, 4H), 7.56 (d, 1H, J ) 2.0 Hz), 7.01 (dd, 1H, J ) 8.7 Hz, 2.2 Hz), 3.83 (s, 3H). 13C NMR (DMSO-d6, 100 MHz): δ 157.5, 140.3, 138.9, 134.2, 133.0, 132.0, 127.6, 124.6, 120.9, 120.4, 114.9, 105.2, 55.5. APCI-MS: m/z 319.1/321.0 (98 /100%) [M + H]+. 6-Methoxy-2-(4-bromophenyl)-3-bromobenzo[b]thiophene (7). Compound 6 (3.13 g, 9.87 mmol) was suspended in dried DCM (70 mL), and N-bromoacetamide (1.45 g, 10.05 mmol) was added with stirring. The reaction mixture was stirred at room temperature for 2 h, diluted with another 100 mL of DCM, and washed with H2O and brine, and the organic phase was separated and concentrated. Product was obtained after recrystallization, using ethyl acetate, as slightly brown solid (3.70 g, 95%). 1H NMR (DMSOd6, 400 MHz): δ 7.66-7.76 (m, 6H), 7.17 (dd, 1H, J ) 8.8 Hz, 2.3 Hz), 3.87 (s, 3H). 13C NMR (DMSO-d6, 100 MHz): δ 158.3, 138.5, 133.6, 132.2, 131.9, 131.6, 131.0, 123.9, 122.3, 115.8, 105.4, 104.3, 55.7. APCI-MS: m/z 398.8 [M + H]+. 6-Methoxy-2-(4-bromophenyl)-3-bromobenzo[b]thiophene Soxide (8). Compound 7 (2.6 g, 6.53mmol) was dissolved in dichloromethane (30 mL), TFA (25 mL) was added dropwise with stirring, then 1.4 mL of H2O2 (30%) was added, and the reaction mixture was stirred at room temperature for 2 h. Sodium bissulfide (500 mg) in 5 mL of H2O was added and stirred vigorously for 15 min to quench the reaction. Most of the solvent was removed under reduced pressure, and the residue was diluted with dichloromethane and carefully washed with saturated aqueous NaHCO3 solution. After concentration, the crude product was purified by column chromatography, eluting with 6:1 hexanes/ethyl acetate. Product was obtained as slightly yellow solid (1.75 g, 65%). 1H NMR (DMSO-d6, 400 MHz): δ 7.67-7.80 (m, 5H), 7.62 (d, 1H, J ) 8.5 Hz), 7.31 (d, 1H, J ) 8.5 Hz), 3.91 (s, 3H). 13C NMR (DMSOd6, 100 MHz): δ 161.2, 144.4, 143.8, 132.1, 131.0, 129.2, 128.9, 125.6, 123.1, 121.9, 118.7, 112.4, 56.3. APCI-MS: m/z 414.9 (100%) [M + H]+. 6-Methoxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-(4-bromophenyl) Benzo[b]thiophene S-oxide (9). The side chain phenol compound (1.80 g, 8.13 mmol) was dissolved in anhydrous DMF (50 mL), NaH (328 mg, 60%, dispersed in oil) was added in three portions within 20 min, and then bromide 8 (3.3 g, 8.01mmol) was added. The reaction mixture was stirred at room temperature for 1 h and diluted with 200 mL of ethyl acetate, and the resulting solution was washed with water. The aqueous phase was back extracted with 50 mL of DCM, combined with the ethyl acetate solution, and then dried with anhydrous MgSO4. Solvent was removed under high vacuum to get a yellow solid that was pure enough for the next step reaction (4.2 g, 94%). Analytical sample was obtained by PTLC (5:1 DCM/MeOH). 1H NMR (DMSO-d6, 400 MHz): δ 7.54 (m, 1H), 7.59-7.70 (m, 4H), 7.06-7.11 (m, 3H), 7.01(d, 1H, J ) 8.5 Hz), 6.88 (d, 2H, J ) 8.8 Hz), 3.98 (t, 2H, J ) 5.6 Hz), 3.86 (s, 3H), 2.59 (t, 2H, J ) 5.6 Hz), 2.39 (br s, 4H), 1.42-1.52 (m, 4H), 1.32-1.42 (m, 2H). 13C NMR (DMSO-

Benzothiophene SERMs

d6, 75 MHz): 160.8, 155.1, 150.1, 148.4, 144.8, 131.9, 129.8, 129.6, 128.8, 125.3, 123.9, 121.6, 118.1, 118.0, 115.7, 112,8, 66.0, 57.3, 56.1, 54.4, 25.6, 23.9. APCI-MS: m/z 554.2/556.1 (100/98%) [M + H]+. 6-Methoxy-3-{4-[2-(1-piperidinyl) ethoxy]phenoxy}-2-(4-bromophenyl) Benzo[b]thiophene (10). Compound 9 (4.0 g, 7.2 mmol) was dissolved in DCM (25 mL), and 1 M HCl/ether solution (15 mL) was added and stirred for 1 h. All solvent was removed under reduced pressure. The residue was dissolved in anhydrous THF (70 mL), TMSCl (9.2 mL, 72 mmol) and Ph3P (7.0 g, 26.7 mmol) were added, the reaction mixture was refluxed for 8h. Most of the solvent was removed, residue was diluted with 200 mL ethyl acetate, washed with saturated aqueous NaHCO3 solution, concentrated, crude product was purified by column chromatography, eluting with 30:1 DCM/MeOH containing 3‰ HOAc to remove the excess of Ph3P and Ph3PO, then eluting with 20:1 DCM/MeOH to get the product (3.3 g, 85%). 1H NMR (DMSO-d6, 400 MHz): δ 7.61-7.68 (m, 5H), 7.21 (d, 1H, J)8.8 Hz), 6.95 (dd, 1H, J)8.8 Hz, 2.3 Hz), 6.86-6.89 (m, 4H), 3.95 (t, 3H, J) 5.9 Hz), 3.83 (s, 3H), 2.59 (t, 3H, J) 5.9 Hz), 2.38 (bs, 4H), 1.43-1.48 (m, 4H), 1.35-1.36 (m, 2H); 13C NMR (DMSO-d6, 75 MHz): δ 158.2 154.0, 150.7 141.0, 136.8 132.1 130.9, 128.7, 126.9, 124.3, 122.0, 121.0, 116.2, 115.6, 115.1, 106.1, 65.9, 57.4, 55.6, 54.4, 25.5, 23.9; HRMS calcd. for C28H29NO3SBr 538.1052 [M+H]+, found 538.1072. 6-Hydroxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-(4-bromophenyl) benzo[b] thio phene (4). Compound 10 (300 mg, 0.56 mmol) was dissolved in DCM (3 mL), 1N HCl in diethyl ether (3 mL) was added, the mixture was stirred at room temperature for 30 min. All solvents was removed and the obtained slightly yellow foam was redissolved in 15 mL DCM, the flask was filled with argon, BF3 dimethyl sulfide complex (2.2 mL) was added, the resulting mixture was stirred at room temperature for 5h. The reaction was diluted with 30 mL ethyl acetate, and washed with saturated NaHCO3 aqueous solution, organic phase was separated and dried by anhydrous MgSO4. The crude product was purified by column chromatography (15:1 DCM/MeOH), the product was obtained as slight yellow solid (240 mg, 82%). 1H NMR (DMSOd6, 400 MHz): δ 9.92 (s, 1H), 7.58-7.67 (m, 4H), 7.30 (d, 1H, J)1.9 Hz), 7.14 (d, 1H, J)8.7 Hz), 6.85 (bs, 4H), 6.82 (dd, 1H, J)8.7 Hz, 1.9 Hz), 3.97 (t, 2H, J) 5.9 Hz), 2.59 (t, 2H, J) 5.8 Hz), 2.38(m, 4H), 1.43-1.48 (m, 4H), 1.32-1.38 (m, 2H) ; 13C NMR (DMSO-d6, 100 MHz): δ 156.4, 154.0, 150.7, 141.2, 136.7, 132.0 , 131.1 , 128.6, 125.9, 123.1, 122.2, 120.8, 116.2, 115.6, 115.2, 108.0, 65.8, 57.3, 54.3, 25.4, 23.8 ; HRMS calcd. for C27H27NO3SBr 524.0895 [M+H]+, found 524.0887. 6-Methoxy-3-{4-[2-(1-piperidinyl) ethoxy]phenoxy}-2-phenyl benzo[b] thiophene (11). Compound 9 (270 mg, 0.47 mmol) was dissolved in anhydrous THF (5 mL), LiAlH4 (55 mg, 1.45 mmol) was added, the reaction mixture was refluxed for 6h. The reaction was quenched by adding 2N aqueous NaOH solution (0.5 mL), then diluted with 30 mL ethyl acetate, more NaOH was added until most of the white precipitate dissolved. The organic phase was separated and concentrated, the residue was purified by column chromatography (50:3 DCM/MeOH), product was obtained as white solid (180 mg, 83%). 1H NMR (CDCl3, 400 MHz): δ 7.76 (d, 2H, J) 7.4 Hz), 7.25-7.37 (m, 5H), 6.86-6.91 (m, 3H), 6.79-6.81 (m, 2H), 4.05 (t, 2H, J) 6.0 Hz), 3.87 (s, 3H), 2.76 (t, 2H, J) 6.0 Hz), 2.52 (bs, 4H), 1.59-1.65 (m, 4H), 1.42-1.48 (m, 2H); 13C NMR (CDCl3, 75 MHz): δ158.2, 154.2, 151.9, 141.0, 137.4, 132.6, 128.9, 128.2, 127.8, 127.6, 126.7, 122.7, 116.6, 115.7, 114.6, 105.5, 66.5, 58.1, 55.8, 55.2, 26.0, 24.8; HRMS calcd. for C28H30NO3S 460.1946 [M+H]+, found 460.1935. 6-Hydroxy-3-{4-[2-(1-piperidinyl) ethoxy]phenoxy}-2-phenyl benzo[b] thiophene (12). Compound 11 (150 mg 0.33 mmol) was dissolved in DCM (5 mL), 1N HCl in diethyl ether (1 mL) was added, the mixture was stirred at room temperature for 30 min. All solvents was removed and the residue was redissolved in 8 mL DCM, the flask was filled with argon, BF3 dimethyl sulfide complex (1.7 mL) was added, the resulting mixture was stirred at room temperature for 5h. The reaction was diluted with 30 mL ethyl acetate, and washed with saturated NaHCO3 aqueous solution,

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2689

the organic phase was separated and dried by anhydrous MgSO4. The crude product was purified by column chromatography (12:1 to 7:1 DCM/MeOH), the product was obtained as white solid (95 mg, 65%). 1H NMR (Acetone-d6, 400 MHz): δ 7.75-7.77 (m, 2H), 7.39 (m, 2H), 7.35 (d, 1H, J)1.8 Hz), 7.22-7.32 (m, 2H), 6.85-6.92 (m, 5H), 4.07 (t, 3H, J) 5.8 Hz), 2.77 (t, 3H, J) 5.8 Hz), 2.57 (bs, 4H), 1.54-1.58 (m, 4H), 1.42-1.44 (m,2H); 13C NMR (Acetone-d6, 75 MHz): δ 157.1, 155.2, 152.5, 141.9, 138.2, 133.4, 129.7, 128.6, 128.0, 127.9, 126.2, 123.2, 117.1, 116.4, 115.7, 108.8, 66.9, 58.5, 55.6, 26.4, 24.7; HRMS calcd. for C27H28NO3S 446.1790 [M+H]+, found 446.1785. 6-Hydroxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-(4-methoxyphenyl) benzo[b] thiophene (1, Arzoxifene). Compound 4 (65 mg, 0.12 mmol), CuI (24 mg, 0.12 mmol) were added into an argonflushed flask, DMF (0.7 mL), anhydrous MeOH (1.5 mL) and ethyl acetate (40 µL ,0.4 mmol) were added, after the addition of NaOMe (530 mg, 9.9 mmol, concentration 4.5 M), the reaction mixture were heated at 110 °C for 8h. The reaction mixture was diluted with 60 mL ethyl acetate, neutralized with aqueous 5N acetic acid, then washed with brine, the organic phase was separated, after concentration, the residue was purified by column chromatography (12:1 DCM/MeOH), the product was obtained as white solid (48 mg, 85%). 1H NMR (Acetone-d6, 400 MHz): δ 7.67-7.69 (m, 2H), 7.32 (d, 1H, J) 2.0 Hz), 7.21 (d, 1H, J)8.6 Hz), 6.94-6.96 (m, 2H), 6.86-6.88 (m, 5H), 4.14 (t, 2H, J) 5.8 Hz), 3.80 (s, 3H), 2.92 (t, 2H, J)5.8 Hz), 2.73 (m, 4H), 1.56-1.66 (m, 4H), 1.431.45 (m, 2H); 13C NMR (Acetone-d6, 75 MHz): δ 160.3, 156.8, 154.9, 152.7, 140.7, 137.7, 129.4, 128.1, 126.3, 125.8, 122.8, 117.1, 116.5, 115.6, 115.1, 108.9, 66.3, 58.2, 55.6, 55.4, 25.9, 24.3; HRMS calcd. for C28H30NO4S 476.1896 [M+H]+, found 476.1893. 6-Hydroxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-[(4-methanesulfonyl)phenyl] benzo[b] thiophene (13). Compound 4 (50 mg, 0.095mmol), CH3SO2Na (78 mg, 0.76 mmol), L-proline (8.7 mg, 0.076 mmol), NaOH (6 mg, 0.15 mmol), CuI (14.5 mg, 0.076 mmol) were added to a 5 mL flask, then filled with argon, anhydrous DMSO (1.5 mL) was added, the resulting reaction mixture was stirred at 110 °C for 13 h. The reaction mixture was partitioned between ethyl acetate (30 mL) and water (5 mL), the undissolved solid was filtered off, solid was washed with another 10 mL ethyl acetate. The organic phase was combined, separated and concentrated, crude product was purified by column chromatography (DCM/MeOH 20:1), product was obtained as slightly green foam (45 mg, 90%). 1H NMR (Acetone-d6, 400 MHz): δ 7.93-8.01 (m, 4H), 7.38 (d, 1H, J)1.9 Hz), 7.27 (d, 1H, J)8.7 Hz), 6.80-6.93(m, 5H), 4.03 (t, 2H, J) 5.9 Hz), 3.12 (s, 3H), 2.68 (t, 2H, J)5.9 Hz), 2.48 (m, 4H), 1.50-1.54 (m, 4H), 1.39-1.41 (m, 2H) 13C NMR (Acetone-d6, 75 MHz): δ 157.8, 155.6, 152.1, 144.1, 140.7, 138.9, 138.4, 128.8, 128.3, 127.6, 124.0, 123.9, 117.3, 116.5, 116.2, 108.9, 67.2, 58.6, 55.6, 44.3, 26.7, 24.9. HRMS calcd. for C28H30NO5S2 524.1565 [M+H]+, found 524.1575 6-Methoxy-3-{4-[2-(1-piperidinyl) ethoxy]phenoxy}-2-(4-aminophenyl) benzo[b] thiophene (14). Compound 10 (150 mg, 0.278 mmol), NaN3 (217 mg, 3.36 mmol), L-proline (28 mg, 0.25 mmol), NaOH (11.1 mg, 0.27 mmol), CuI (47 mg, 0.25 mmol) were added to a 10 mL flask, then filled with argon, a mixture of 4 mL DMSO/2 mL EtOH was added by syringe, the resulting reaction mixture was stirred at 110 °C for 10h. The reaction was partitioned between ethyl acetate and water, the undissolved solid was filtered off. The organic phase was separated and concentrated, crude product was purified by column chromatography (DCM/MeOH), product was obtained as slightly green syrup (98 mg, 74%). 1H NMR (DMSOd6, 400 MHz): δ 8.31 (s,1H), 7.51 (d, 1H, J)2.0 Hz), 7.37 (d, 2H, J)8.4 Hz), 7.12 (d, 1H, J)8.7 Hz), 6.81-6.91 (m, 5H), 6.55 (d, 2H, J)8.4 Hz), 5.40 (s, 2H), 3.95 (t, 2H, J) 5.7 Hz), 3.80 (s, 3H), 2.59 (t, 2H, J)5.6 Hz), 2.39 (bs, 4H), 1.45-1.50 (m, 4H), 1.351.40 (m, 2H); 13C NMR (DMSO-d6, 100 MHz): δ 157.3, 153.7, 151.0, 148.9, 137.7, 135.4, 127.9, 127.7, 127.4, 120.9, 118.8, 115.9, 115.5, 114.3, 113.8, 106.0, 65.8, 57.4, 55.5, 54.3, 25.5, 23.9; HRMS calcd. for C28H31N2O3S 475.2055 [M+H]+, found 475.2045. 6-Hydroxy-3-{4-[2-(1-piperidinyl) ethoxy]phenoxy}-2-(4-aminophenyl) benzo[b] thiophene (15). Compound 14 (250 mg, 0.53

2690

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11

mmol) was dissolved in DCM (10 mL), 1M HCl in ether (3 mL) was added and the mixture was stirred at room temperature for 30 min. All solvents was removed and the residue was redissolved in DCM (12 mL), the flask was filled with argon, BF3 dimethyl sulfide complex (2.5 mL) was added, the resulting mixture was stirred at room temperature for 5h. The reaction was diluted with 30 mL ethyl acetate, and washed with saturated NaHCO3 aqueous solution, the organic phase was separated and dried by anhydrous MgSO4. The crude product was purified by column chromatography (10:1 DCM/MeOH), the product was obtained as slightly yellow solid (175 mg, 72%). 1H NMR (Acetone-d6, 400 MHz): δ 7.47 (d, 2H, J)8.68 Hz), 7.28 (d, 1H, J)1.86 Hz), 7.16 (d, 1H, J)8.62 Hz), 6.38-6.91 (m, 5H), 6.66 (d, 2H, J ) 6.89 Hz), 4.89 (bs, 2H), 4.21 (t, 2H, J) 5.45 Hz), 3.19 (t, 2H, J) 5.36 Hz), 3.00 (m, 4H), 1.711.76 (m, 4H), 1.52-1.56 (m, 2H); 13C NMR (Acetone-d6, 100 MHz): δ 156.3, 155.1, 152.6, 139.5, 137.1, 129.0, 128.4, 127.7, 122.4, 121.5, 116.9, 116.3, 115.2, 115.1, 108.7, 67.3, 58.7, 55.7, 26.7, 24.9; HRMS calcd. for C27H29N2O3S 461.1899 [M+H]+, found 461.1901. 6-Hydroxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-[4-(2chloroacetamide)phenyl] benzo[b] thiophene (16). Compound 15 (46 mg, 0.1 mmol), pyridine (80 µL) were dissolved in anhydrous CH2Cl2 (1.5 mL) and cooled in ice bath, chloroacetyl chloride (6 µL) was added by a syringe. The reaction mixture was gradually warmed up to room temperature and stirred overnight. Reaction mixture was diluted with CH2Cl2 (25 mL), washed with saturated aqueous NaHCO3 solution. The organic phase was separated and concentrated, crude product was purified by flash column chromatography (AcOEt/CH2Cl2/MeOH 30:25:8). The product was obtained as slightly yellow foam (31 mg, 57%). 1H NMR (Acetoned6, 400 MHz): δ 9.56 (s, 1H), 7.74-7.69 (m, 4H), 7.33 (d, 1H, J) 2.0 Hz), 7.22 (d, 1H, J) 8.65 Hz), 6.91-6.83 (m, 5H), 4.24 (s,1H), 4.01 (t, 2H, J) 5.96 Hz), 2.66 (t, 2H, J)5.96 Hz), 2.46 (m, 4H), 1.53-1.49 (m, 4H), 1.40-1.35 (m, 2H); 13C NMR (Acetone-d6, 75 MHz): δ 165.5, 157.1, 155.4, 152.5, 141.7, 139.1, 138.0, 129.2, 128.6, 128.1, 125.9, 123.2, 120.7, 117.2, 116.5, 115.8, 108.9, 67.4, 58.8, 55.8, 44.2, 26.8, 25.1; HRMS calcd. for C29H30N2O4SCl 537.1615 [M+H]+, found 537.1595. 6-Methoxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-{4-[(2chloroethyl) amino] phenyl} benzo[b] thiophene (17). Compound 14 (60 mg 0.12 mmol) was dissolved in 2 mL methanol, chloroacetaldehyde (30 µL 45% aqueous solution, 0.16 mmol), NaBH3CN (15 mg, 0.17 mmol), 6N HCl in methanol (30 µL) were added, the reaction mixture was stirred at room temperature for 3 days, solvent was removed under reduced pressure, residue was purified by column chromatography (DCM/MeOH 20:1 containing 1‰ NH3‚H2O) to get the product (25 mg, 37.3%). 1H NMR (Acetoned6, 400 MHz): δ 7.55(d, 2H, J)8.7 Hz), 7.43(d, 1H, J)2.1 Hz), 7.21 (d, 1H, J)8.7 Hz), 6.84-6.9 3 (m, 5H), 6.69 (d, 2H, J)8.7 Hz), 5.50 (t, 1H, J) 5.96 Hz), 4.01 (t, 2H, J) 5.9 Hz), 3.86 (s, 3H), 3.73 (t, 2H, J)6.2 Hz), 3.52-3.55 (m, 2H), 2.66 (t, 2H, J)6.0 Hz), 2.46 (m, 4H), 1.49-1.55 (m, 4H), 1.38-1.41 (m, 2H); 13C NMR (Acetone-d6, 75 MHz): δ 158.8, 155.2, 152.6, 148.8, 139.6, 137.2, 129.2, 129.2, 128.3, 122.3, 121.7, 116.9, 116.4, 115.1, 113.4, 106.5, 67.3, 58.7, 55.9, 55.7, 45.8, 43.8, 26.7, 24.9; HRMS calcd. for C30H34N2O3SCl 537.1979 [M+H]+, found 537.1983. 6-Hydroxy-3-{4-[2-(1-piperidinyl)ethoxy]phenoxy}-2-{4-[(2chloroethyl) amino] phenyl} benzo[b] thiophene (18). Compound 17 (20 mg, 0.03 7mmol) was dissolved in DCM (1.5 mL), 1N HCl in diethyl ether (0.3 mL) was added, the mixture was stirred at room temperature for 30 min. All solvents was removed and the obtained brown foam was redissolved in DCM (1.5 mL), the flask was filled with argon, BCl3 dimethyl sulfide complex (0.3 mL) was added, the resulting mixture was stirred at 70 °C for 10 h. The reaction was diluted with 10 mL DCM, and washed with saturated NaHCO3 aqueous solution, the organic phase was separated and dried by anhydrous MgSO4. Crude product was purified by column chromatography (15:1 DCM/MeOH). Product was obtained as slight yellow solid (15 mg, 77%). 1H NMR (Acetone-d6, 400 MHz): δ 7.54 (d, 2H, J)8.7 Hz), 7.29 (d, 1H, J)1.9 Hz), 7.16 (d, 1H, J)8.6 Hz), 6.83-6.89 (m, 5H), 6.69 (d, 2H, J)8.7 Hz), 5.49 (t, 1H, J)

Qin et al.

5.9 Hz), 4.10 (t, 2H, J) 5.8 Hz), 3.73 (t, 2H, J)6.1 Hz), 3.513.56 (m, 2H), 2.79 (t, 2H, J)5.7 Hz), 2.60 (m, 4H), 1.56-1.62 (m, 4H), 1.40-1.46 (m, 2H); 13C NMR (Acetone-d6, 75 MHz): δ 156.5, 155.1, 152.8, 148.8, 139.7, 137.3, 129.3, 128.5, 127.5, 122.5, 121.9, 117.1, 116.5, 115.4, 113.4, 108.9, 66.9, 58.5, 55.5, 45.9, 43.9, 26.3, 24.7; HRMS calcd. for C29H32N2O3SCl 523.1822 [M+H]+, found 523.1809. Estrogen Receptor Binding Assays. ER competitive binding assay with [3H]-estradiol. The assay was slightly modified from the original protocol.49 Twenty-four hours before the assay, 50% v/v hydroxyapatite (HAP) slurry was prepared using 10 g hydroxylapatite/60 mL of TE buffer (50 mM Tris-Cl, 1 mM EDTA, pH 7.4) and stored at 4 °C. ER binding buffer (10 mM Tris, 10% glycerol, 2 mM dithiothrietol, 1 mg/mL bovine serum albumin, pH 7.5), ERR (40 mM Tris, 100 mM KCl, pH 7.5) and ERβ (40 mM Tris, pH 7.5) wash buffers were prepared subsequently. The reaction mixture consisted of 5µL of test samples in DMSO, 5µL of pure human recombinant diluted ERR or ERβ (0.5pmol) in ER binding buffer, 5µL of “Hot Mix” (400nM, prepared fresh using 3.2µL of 25µM, 83Ci/mM 3H-estradiol, 98.4µL of ER binding buffer), and 85µL ER binding buffer. The incubations were performed at room temperature for 2 h or at 4 °C overnight, then 100µL of 50% HAP slurry was added and the tubes were incubated on ice for 15 min with votexing every 5 min. The appropriate ER wash buffer was added (1 mL), and the tubes were vortexed and then centrifuged at 2000 g for 5 min. The supernatant was discarded and this wash step was repeated three times. The HAP pellet, containing the ligand-receptor complex, was then resuspended in 200 µL of ethanol and transferred to scintillation vials. Cytoscint (4 mL/vial) was added, and the samples were counted using a Beckman (Schaumburg, IL) LS 5801 liquid scintillation counter. The percent inhibition of [3H] estradiol binding to each ER was determined as follows: [1 - (dpmsample - dpmblank)/(dpmDMSO - dpmblank)] × 100. IC50 values were calculated from binding of the sample expressed as a percentage relative to E2 (50 nM, 100%). Relative binding affinity (RBA; relative to E2) was calculated from IC50(E2)/IC50(sample). The samples were assayed in triplicate at at least five concentrations. Antioxidant Activities of SERMs. Antioxidant activities of SERMs were monitored using DPPH assay. The 0.2 mM solution of DPPH in methanol and 1.2 mM of SERMs in DMSO were mixed in a 1 mL spectrophotometer cell. The maximum volume of DMSO used in experiments was less than 4% compare to methanol. Kinetic traces were obtained at 515 nm using a HP8452A diode array spectrometer. Observed rate constants (kobs) were obtained from first-order fitting of the corresponding kinetic traces. Excellent pseudo first-order kinetics were observed for DPPH degradation in the presence of monophenolic SERMs (Figure 5), although for 2 and 15, the curve shape suggested more complex behavior. Computational Methods. DFT molecular orbital calculations were performed using Spartan 4.0 for Windows (Wavefunction Inc., CA) using minimum energy conformations obtained at the semiempirical AM1 level. All molecular modeling studies were performed on an SGI computer with the Sybyl 7.2 software packages. The coordinates for the estrogen receptor alpha (ERR) LBD were extracted from the cocrystal structure data of the complex between ERR LBD and raloxifene (PDB code: 1ERR). The coordinates for the estrogen receptor beta (ERβ) LBD were extracted from the cocrystal structure data of the complex between ERβ and tamoxifen (PDB code: 2FSZ). The active site was designated to consist of the amino acid residues within a radius of 4.5 Å from the original ligand raloxifene or tamoxifen. According to ERR (ERβ in parenthesis) reference numbering, Asp351(303), Glu353(305), Arg394(346), and His524(475) were set as a core subpocket. The following FlexX-Pharm settings were used to restrict the binding of the ligands to the raloxifene/tamoxifen binding site: Asp351(303) is an optional hydrogen-bond acceptor, and a spatial constraint of r ) 3 Å is set around carboxylate oxygens of Asp351(303) and the ligand’s nitrogen. Arzoxifene and its 4′-position modified analogs were screened. After ligand docking was performed by the FlexX and Flex-Pharm

Benzothiophene SERMs

modules in Sybyl, the 30 best poses were selected for each ligand and saved for analysis by CScore. We used a combination of several functions and the criterion of consensus g4 to select the best pose for each ligand. The binding modes of the docked ligands were found to be consistent with those expected for compounds that are structurally related to raloxifene and tamoxifen. The selection of the binding poses was based on FlexX score. Scoring and ranking of the poses using the CScore scoring function performed reasonably well for the ligands able to dock within the LBD cavity. Bulkier 4′-substituents failed to dock inside the cavity. For these ligands, manual docking was designed based on the top scoring DMA pose. Furthermore, the most meaningful pose or manually docked pose for each ligand was co-minimized with the receptor LBD by employing the Powell method with initial Simplex optimization (20 steps) using the Tripos force field, Gasteiger-Huckel charges, dielectric constant ) 1, and nonbonding cutoff of 8 Å, until a convergence of 0.05 kcal/mol·Å was reached. The minimized poses were re-evaluated with CScore functions. To examine the second, low-affinity reported binding site for tamoxifen in ERβ, a second screening for arzoxifene analogs was devised based on PBD ID: 2FSZ. Because most of the interaction of the low affinity binding site are based on van der Waals contacts, the second active site was designated to consist of amino acids with a radius of 5.5 Å from tamoxifen, with core subpocked residues such as Leu306, Met309, Ile310 Val 328, Leu331, Glu332, and Trp335 according to ERβ 2FSZ reference numbering. The binding resulted in weak, nonspecific association of hydrophobic 4′substituents, with the hydrophobic groove of the coactivator recognition surface. Energy Evaluation. To compare the energetic interactions of the arzoxifene series with the estrogen receptor ERR vs raloxifene interaction with ERR, the following equation was used ∆Gbind ) (GER-analog complex - Ganalog) (GER-raloxifene complex - GER-raloxifene) Solute energy was evaluated for the lowest energy conformer in the same method as mentioned above for the ERR-analog complex.

Acknowledgment. This work was supported by NIH Grants CA102590 and CA79870. References (1) Jordan, V. C. Tamoxifen: A most unlikely pioneering medicine. Nat. ReV. Drug DiscoVery 2003, 2, 205-213. (2) Ariazi, E. A.; Ariazi, J. L.; Cordera, F.; Jordan, V. C. Estrogen receptors as therapeutic targets in breast cancer. Curr. Top. Med. Chem. 2006, 6, 181-202. (3) Rossouw, J. E.; Anderson, G. L.; Prentice, R. L.; LaCroix, A. Z.; Kooperberg, C.; Stefanick, M. L.; Jackson, R. D.; Beresford, S. A.; Howard, B. V.; Johnson, K. C.; Kotchen, J. M.; Ockene, J. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results from the Women’s Health Initiative randomized controlled trial. JAMA, J. Am. Med. Assoc. 2002, 288, 321-333. (4) Vogelvang, T. E.; van der Mooren, M. J.; Mijatovic, V.; Kenemans, P. Emerging selective estrogen receptor modulators: Special focus on effects on coronary heart disease in postmenopausal women. Drugs 2006, 66, 191-221. (5) Barrett-Connor, E.; Mosca, L.; Collins, P.; Geiger, M. J.; Grady, D.; Kornitzer, M.; McNabb, M. A.; Wenger, N. K. Effects of raloxifene on cardiovascular events and breast cancer in postmenopausal women. N. Engl. J. Med. 2006, 355, 125-137. (6) Pines, A. Lessons from RUTH. Climacteric 2006, 9, 323-324. (7) Gol, M.; Baris, N.; Guneri, S.; Posaci, C. The effect of raloxifene on cardiac autonomic regulation in osteoporotic women. Am. J. Obstet. Gynecol. 2006, 194, 1249-1254. (8) Suh, N.; Glasebrook, A. L.; Palkowitz, A. D.; Bryant, H. U.; Burris, L. L.; Starling, J. J.; Pearce, H. L.; Williams, C.; Peer, C.; Wang, Y.; Sporn, M. B. Arzoxifene, a new selective estrogen receptor modulator for chemoprevention of experimental breast cancer. Cancer Res. 2001, 61, 8412-8415. (9) Munster, P. N.; Buzdar, A.; Dhingra, K.; Enas, N.; Ni, L.; Major, M.; Melemed, A.; Seidman, A.; Booser, D.; Theriault, R.; Norton, L.; Hudis, C. Phase I study of a third-generation selective estrogen receptor modulator, LY353381.HCL, in metastatic breast cancer. J. Clin. Oncol. 2001, 19, 2002-2009.

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2691 (10) Kim, S. Y.; Suzuki, N.; Laxmi, Y. R.; Shibutani, S. Genotoxic mechanism of tamoxifen in developing endometrial cancer. Drug Metab. ReV. 2004, 36, 199-218. (11) Dowers, T. S.; Qin, Z. H.; Thatcher, G. R.; Bolton, J. L. Bioactivation of selective estrogen receptor modulators (SERMs). Chem. Res. Toxicol. 2006, 19, 1125-1137. (12) Bhat, H. K.; Calaf, G.; Hei, T. K.; Loya, T.; Vadgama, J. V. Critical role of oxidative stress in estrogen-induced carcinogenesis. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3913-3918. (13) Shen, L.; Pisha, E.; Huang, Z.; Pezzuto, J. M.; Krol, E.; Alam, Z.; van Breemen, R. B.; Bolton, J. L. Bioreductive activation of catechol estrogen-ortho-quinones: Aromatization of the B ring in 4-hydroxyequilenin markedly alters quinoid formation and reactivity. Carcinogenesis 1997, 18, 1093-1101. (14) Bolton, J. L.; Yu, L.; Thatcher, G. R. J. Quinoids formed from estrogens and antiestrogens. Methods Enzymol. 2004, 378, 110-123. (15) Liu, H.; Liu, J.; van Breemen, R. B.; Thatcher, G. R. J.; Bolton, J. L. Bioactivation of the selective estrogen receptor modulator desmethylated arzoxifene to quinoids: 4′-Fluoro substitution prevents quinoid formation. Chem. Res. Toxicol. 2005, 18, 162-173. (16) Liu, H.; Bolton, J. L.; Thatcher, G. R. J. Chemical modification modulates estrogenic activity, oxidative reactivity, and metabolic stability in 4′-F-DMA, a new benzothiophene selective estrogen receptor modulator. Chem. Res. Toxicol. 2006, 19, 779-787. (17) Yao, D.; Zhang, F.; Yu, L.; Yang, Y.; van Breemen, R. B.; Bolton, J. L. Synthesis and reactivity of potential toxic metabolites of tamoxifen analogues: Droloxifene and toremifene o-quinones. Chem. Res. Toxicol. 2001, 14, 1643-1653. (18) Liu, J.; Liu, H.; van Breemen, R. B.; Thatcher, G. R. J.; Bolton, J. L. Bioactivation of the selective estrogen receptor modulator acolbifene to quinone methides. Chem. Res. Toxicol. 2005, 18, 174182. (19) Liu, J.; Li, Q.; Yang, X.; van Breemen, R. B.; Bolton, J. L.; Thatcher, G. R. J. Analysis of protein covalent modification by xenobiotics using a covert oxidatively activated tag: raloxifene proof-of-principle study. Chem. Res. Toxicol. 2005, 18, 1485-1496. (20) Yu, L.; Liu, H.; Li, W.; Zhang, F.; Luckie, C.; van Breemen, R. B.; Thatcher, G. R. J.; Bolton, J. L. Oxidation of raloxifene to quinoids: Potential toxic pathways via a diquinone methide and o-quinones. Chem. Res. Toxicol. 2004, 17, 879-888. (21) Zhang, D. D.; Hannink, M. Distinct cysteine residues in Keap1 are required for Keap1-dependent ubiquitination of Nrf2 and for stabilization of Nrf2 by chemopreventive agents and oxidative stress. Mol. Cell. Biol. 2003, 23, 8137-8151. (22) Dinkova-Kostova, A. T.; Holtzclaw, W. D.; Cole, R. N.; Itoh, K.; Wakabayashi, N.; Katoh, Y.; Yamamoto, M.; Talalay, P. Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11908-11913. (23) Evans, D. C.; Watt, A. P.; Nicoll-Griffith, D. A.; Baillie, T. A. Drugprotein adducts: An industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 2004, 17, 3-16. (24) Sporn, M. B. Arzoxifene: A promising new selective estrogen receptor modulator for clinical chemoprevention of breast cancer. Clin. Cancer Res. 2004, 10, 5313-5315. (25) Palkowitz, A. D.; Glasebrook, A. L.; Thrasher, K. J.; Hauser, K. L.; Short, L. L.; Phillips, D. L.; Muehl, B. S.; Sato, M.; Shetler, P. K.; Cullinan, G. J.; Pell, T. R.; Bryant, H. U. Discovery and synthesis of [6-hydroxy-3-[4-[2-(1-piperidinyl)ethoxy]phenoxy]-2-(4-hydroxyphenyl)]benzo[b]thiophene: A novel, highly potent, selective estrogen receptor modulator. J. Med. Chem. 1997, 40, 1407-1416. (26) Zuckerman, S. H.; Bryan, N. Inhibition of LDL oxidation and myeloperoxidase dependent tyrosyl radical formation by the selective estrogen receptor modulator raloxifene (LY139481 HCL). Atherosclerosis 1996, 126, 65-75. (27) Katzenellenbogen, B. S.; Choi, I.; Delage-Mourroux, R.; Ediger, T. R.; Martini, P. G.; Montano, M.; Sun, J.; Weis, K.; Katzenellenbogen, J. A. Molecular mechanisms of estrogen action: Selective ligands and receptor pharmacology. J. Steroid Biochem. Mol. Biol. 2000, 74, 279-285. (28) Benz, C. C.; Atsriku, C.; Yau, C.; Britton, D.; Schilling, B.; Gibson, B. W.; Baldwin, M. A.; Scott, G. K. Novel pathways associated with quinone-induced stress in breast cancer cells. Drug Metab. ReV. 2006, 38, 601-613. (29) Zoete, V.; Rougee, M.; Dinkova-Kostova, A. T.; Talalay, P.; Bensasson, R. V. Redox ranking of inducers of a cancer-protective enzyme via the energy of their highest occupied molecular orbital. Free Radical Biol. Med. 2004, 36, 1418-1423. (30) Hummel, C. W.; Geiser, A. G.; Bryant, H. U.; Cohen, I. R.; Dally, R. D.; Fong, K. C.; Frank, S. A.; Hinklin, R.; Jones, S. A.; Lewis, G.; McCann, D. J.; Rudmann, D. G.; Shepherd, T. A.; Tian, H.;

2692

(31)

(32)

(33) (34)

(35) (36) (37)

(38)

Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 Wallace, O. B.; Wang, M.; Wang, Y.; Dodge, J. A. A selective estrogen receptor modulator designed for the treatment of uterine leiomyoma with unique tissue specificity for uterus and ovaries in rats. J. Med. Chem. 2005, 48, 6772-6775. Geiser, A. G.; Hummel, C. W.; Draper, M. W.; Henck, J. W.; Cohen, I. R.; Rudmann, D. G.; Donnelly, K. B.; Adrian, M. D.; Shepherd, T. A.; Wallace, O. B.; McCann, D. J.; Oldham, S. W.; Bryant, H. U.; Sato, M.; Dodge, J. A. A new selective estrogen receptor modulator with potent uterine antagonist activity, agonist activity in bone, and minimal ovarian stimulation. Endocrinology 2005, 146, 4524-4535. Brzozowski, A. M.; Pike, A. C.; Dauter, Z.; Hubbard, R. E.; Bonn, T.; Engstrom, O.; Ohman, L.; Greene, G. L.; Gustafsson, J. A.; Carlquist, M. Molecular basis of agonism and antagonism in the oestrogen receptor. Nature 1997, 389, 753-758. Ong, E. B.; Shaw, E.; Schoellmann, G. The identification of the histidine residue at the active center of chymotrypsin. J. Biol. Chem. 1965, 240, 694-698. Mitra, K.; Marquis, J. C.; Hillier, S. M.; Rye, P. T.; Zayas, B.; Lee, A. S.; Essigmann, J. M.; Croy, R. G. A rationally designed genotoxin that selectively destroys estrogen receptor-positive breast cancer cells. J. Am. Chem. Soc. 2002, 124, 1862-1863. Rabasseda, X.; Leeson, P. A.; Castan˜er, J. Arzoxifene hydrochloride. Estrogen receptor modulator, treatment of postmenopausal syndrome, antineoplastic. Drugs Future 1999, 24, 599-604. David, E.; Perrin, J.; Pellet-Rostaing, S.; Fournier dit Chabert, J.; Lemaire, M. Efficient access to 2-aryl-3-substituted benzo[b]thiophenes. J. Org. Chem. 2005, 70, 3569-3573. Samanen, J. M.; Brandeis, E. The p-(methylsulfinyl)benzyl group: A trifluoroacetic acid (TFA)-stable carboxyl-protecting group readily convertible to a TFA-labile group. J. Org. Chem. 1988, 53, 561569. Pinney, K. G.; Carlson, K. E.; Katzenellenbogen, J. A. Nonsteroidal estrogens bearing acyl azide functions: Potential electrophilic and photoaffinity labeling agents for the estrogen receptor. Steroids 1992, 57, 222-232.

Qin et al. (39) Capdeville, P.; Maumy, M. Esters are effective co-catalysts in coppercatalyzed methanolysis of aryl bromides. Tetrahedron Lett. 1993, 34, 1007-1010. (40) Zhu, W.; Ma, D. Synthesis of aryl sulfones via L-proline-promoted CuI-catalyzed coupling reaction of aryl halides with sulfinic acid salts. J. Org. Chem. 2005, 70, 2696-2700. (41) Zhu, W.; Ma, D. Synthesis of aryl azides and vinyl azides via prolinepromoted CuI-catalyzed coupling reactions. Chem. Commun. 2004, 888-889. (42) Mann, J.; Hasse-Held, M.; Springer, C. J.; Bagshawe, K. D. Synthesis of an N-mustard prodrug. Tetrahedron 1990, 46, 5377-5382. (43) Williard, P. G.; Fryhle, C. B. Boron trihalide-methyl sulfide complexes as convenient reagents for dealkylation of aryl ethers. Tetrahedron Lett. 1980, 21, 3731-3734. (44) Obourn, J. D.; Koszewski, N. J.; Notides, A. C. Hormone- and DNAbinding mechanisms of the recombinant human estrogen receptor. Biochemistry 1993, 32, 6229-6236. (45) Ewing, T. J.; Makino, S.; Skillman, A. G.; Kuntz, I. D. DOCK 4.0: search strategies for automated molecular docking of flexible molecule databases. J. Comput. Aided. Mol. Des. 2001, 15, 411428. (46) Powell, M. Restart procedures for the conjugate gradient method. Mathematical Programming. Mathematical Programming 1977, 12, 241-254. (47) Gasteiger, J.; Marsili, M. Iterative partial equalization of orbital electronegativity: a rapid access to atomic charges. Tetrahedron 1980, 36, 3219-3222. (48) Gasteiger, J.; Marsili, M. Prediction of proton magnetic resonance shifts: the dependence on hydrogen charges obtained by iterative partial equalization of orbital electronegativity. Org. Magn. Reson. 1980, 15, 353-360. (49) Wang, Y.; Chirgadze, N. Y.; Briggs, S. L.; Khan, S.; Jensen, E. V. et al. A second binding site for hydroxytamoxifen within the coactivator-binding groove of estrogen receptor beta. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 9908-9911.

JM070079J